What do you think?
Rate this book
AI Agents and Applications: With LangChain, LangGraph, and MCP
Build intelligent LLM-powered applications with agentic workflows and tool-based agents.
AI-powered applications are rapidly becoming the new normal. Personal productivity assistants, coding agents, smarter search, and automated reporting tools are popping up everywhere. The LangChain ecosystem, and standards like MCP are driving this new gold rush. This book helps you claim your spot.
In "AI Agents and With LangChain, LangGraph and MCP," you’ll
Prompt and context engineering for accurate, hallucination-resistant systems
Advanced RAG for summarization, semantic search, and reliable Q&A
Structured, multi-step agentic workflows with LangGraph
Tool-based agents that adapt in real time
Multi-agent systems for complex, real-world tasks
MCP integration to expose, compose, and consume plug-and-play tools"AI Agents and Applications" is your hands-on guide to creating real, production-ready language model solutions. With LangChain and LangGraph, you’ll orchestrate powerful agentic workflows and build dynamic tool-based agents that can search, summarize, reason, and act. You’ll move from essential prompt engineering to advanced Retrieval Augmented Generation (RAG), and finally to deploying multi-agent systems using modern integration standards like the Model Context Protocol (MCP).
About the
Along the way you’ll build concrete applications—summarization and Q&A engines, context-aware chatbots with memory, and tool-using AI agents that orchestrate multi-step workflows with branching logic. For the examples, the book uses Python, LangChain, LangGraph, and LangSmith, but you’ll be able to generalize to other frameworks.
About the
Roberto Infante is an AI innovator with deep FinTech experience, working for a London-based hedge fund. He specializes in building agentic systems for both plain vanilla and exotic quantitative analysis.
AI-powered applications are rapidly becoming the new normal. Personal productivity assistants, coding agents, smarter search, and automated reporting tools are popping up everywhere. The LangChain ecosystem, and standards like MCP are driving this new gold rush. This book helps you claim your spot.
In "AI Agents and With LangChain, LangGraph and MCP," you’ll
Prompt and context engineering for accurate, hallucination-resistant systems
Advanced RAG for summarization, semantic search, and reliable Q&A
Structured, multi-step agentic workflows with LangGraph
Tool-based agents that adapt in real time
Multi-agent systems for complex, real-world tasks
MCP integration to expose, compose, and consume plug-and-play tools"AI Agents and Applications" is your hands-on guide to creating real, production-ready language model solutions. With LangChain and LangGraph, you’ll orchestrate powerful agentic workflows and build dynamic tool-based agents that can search, summarize, reason, and act. You’ll move from essential prompt engineering to advanced Retrieval Augmented Generation (RAG), and finally to deploying multi-agent systems using modern integration standards like the Model Context Protocol (MCP).
About the
Along the way you’ll build concrete applications—summarization and Q&A engines, context-aware chatbots with memory, and tool-using AI agents that orchestrate multi-step workflows with branching logic. For the examples, the book uses Python, LangChain, LangGraph, and LangSmith, but you’ll be able to generalize to other frameworks.
About the
Roberto Infante is an AI innovator with deep FinTech experience, working for a London-based hedge fund. He specializes in building agentic systems for both plain vanilla and exotic quantitative analysis.
- GenresArtificial Intelligence
448 pages, Paperback
Published March 17, 2026
Ratings & Reviews
Friends & Following
Create a free account to discover what your friends think of this book!
Community Reviews
Displaying 1 - 10 of 10 reviews
March 20, 2026
I started this book expecting to not learn much, as I have written many GenAi apps, but, I was surprised with some of the ideas. What I found interesting is that I am used to certain patterns of how to write my agents, but I found it fascinating to see other ideas, for example, something as simple as rewriting the query from a user by having the LLM rewrite it to make it either match better for the purpose of the agent, or it may have been too specific and it may be better if it was changed to be more general. It isn't just the ideas I found interesting, but for many of these ideas they give links to papers that had introduced or added more for those that want to go more into it.
This book is very geared toward vector databases (db) and libraries and graphing db, but and OpenAI, with no mention of Ollama many of the concepts can be useful regardless and people can learn about other options easily enough.
I was thrilled to see information about guardrails to help protect the system and to ensure what is returned to not violate rules, and there is quite a bit of discussion on MCP, where agents can use other agentic systems, and how to call tools, so the agents can be more flexible and use many different types of tools, so, for example, the query may want to use hotel reservation system, airline reservation system and/or find travel guides, without hardcoding the paths, but making it more flexible.
The book starts with various approaches on how to create a GenAI application, and then creating agents, then multi-agent, then MCP and guardrails.
This is a great book, but as was mentioned this isn't a book of recipes, but a way to learn how to write the system.
Using LangGraph and LangSmith is very useful to help make agents developed more effectively.
This book is very geared toward vector databases (db) and libraries and graphing db, but and OpenAI, with no mention of Ollama many of the concepts can be useful regardless and people can learn about other options easily enough.
I was thrilled to see information about guardrails to help protect the system and to ensure what is returned to not violate rules, and there is quite a bit of discussion on MCP, where agents can use other agentic systems, and how to call tools, so the agents can be more flexible and use many different types of tools, so, for example, the query may want to use hotel reservation system, airline reservation system and/or find travel guides, without hardcoding the paths, but making it more flexible.
The book starts with various approaches on how to create a GenAI application, and then creating agents, then multi-agent, then MCP and guardrails.
This is a great book, but as was mentioned this isn't a book of recipes, but a way to learn how to write the system.
Using LangGraph and LangSmith is very useful to help make agents developed more effectively.
March 18, 2026
Most books in this space either oversimplify or drown you in code without explaining the reasoning behind it. This one does neither. The author builds from prompt engineering through RAG pipelines to full multi-agent orchestration in a way that actually makes sense as a progression — each concept earns its place before the next one arrives.
The RAG coverage is thorough and honest — ingestion, chunking, retrieval, and hallucination reduction are treated as real engineering concerns rather than afterthoughts. The LangGraph and MCP sections are where experienced practitioners will find the most value — agentic workflows, tool selection, and multi-step reasoning are handled with genuine depth and clarity.
Based on. technical review of this book it is greatly useful for backend developers, ML engineers, and solution architects building real LLM applications in Python. If you want to understand the why behind the how — this is worth your time.
The RAG coverage is thorough and honest — ingestion, chunking, retrieval, and hallucination reduction are treated as real engineering concerns rather than afterthoughts. The LangGraph and MCP sections are where experienced practitioners will find the most value — agentic workflows, tool selection, and multi-step reasoning are handled with genuine depth and clarity.
Based on. technical review of this book it is greatly useful for backend developers, ML engineers, and solution architects building real LLM applications in Python. If you want to understand the why behind the how — this is worth your time.
Review of advance copy received from Publisher
For those interested in implementing agents and agentic applications (which for most starts with how to use an LLM to take natural language user input, query systems for results based on this input, and then generate natural language responses with answers to provide back to the user) , this is an excellent primer. I really enjoyed this as it breaks down the concepts of creating agents for applications and how they operate in hands on way. It does this using the frameworks of LangChain and LangGraph with hands on examples in python. Half the "reading" is clearly written notebooks that provide sample implementations and the patterns to follow to successfully use agents for RAG. It builds up to multi-agent systems and building and consuming from MCP servers. Great jumpstart for making a beginning with these technologies. I read the nearly complete early access edition. Glad I did.
April 4, 2026
This is a very interesting book for getting started in the world of LLMs, chatbots, RAGs, and agents, as it offers a fairly accessible overview of the current landscape of these tools and technologies. It doesn't require extensive prior knowledge of the subject to get the most out of the book; however, a sufficient technical foundation in programming, databases, and software development in general is necessary. The book introduces LangChain and LangGraph, as well as examples in Python, so knowledge of this programming language is essential. Rather than focusing on the fundamentals, the book emphasizes application building with a very practical approach. Therefore, it's not the best resource if you're looking to learn about AI algorithms, as the focus in this case is elsewhere. The book includes many diagrams and code listings and is easy to read.
March 18, 2026
I was an early reviewer of this book for the publisher.
This book is really good. The author seems to think kind of like me, so I can grok what he is saying pretty well. I like the author's writing style, as I do most Manning authors.
The exercises in the book really challenged me as an approximately low intermediate python (still studying) developer without being too frustrating, and I could often read the code and go, oh, I think see what you are doing, before I started the exercise.
It is definitely not a book for a newbie Python developer, but if you have some experience, don't be afraid to jump in and go for it! I thought it was fun!
This book is really good. The author seems to think kind of like me, so I can grok what he is saying pretty well. I like the author's writing style, as I do most Manning authors.
The exercises in the book really challenged me as an approximately low intermediate python (still studying) developer without being too frustrating, and I could often read the code and go, oh, I think see what you are doing, before I started the exercise.
It is definitely not a book for a newbie Python developer, but if you have some experience, don't be afraid to jump in and go for it! I thought it was fun!
March 18, 2026
This book is very useful for begginers in agents and applications.
The author explains very clearly the foundations of LLMs and guide the reader in the first applications with LLMs, summarization, q&a chatbots, Retrieval-Augmented Generation and AI agents using LangChain, LangGraph, and MCP.
The reader can follow very easily how to build agents because the requirements are very well explained.
The Appendices are very useful if you're not familiar with the settings required in order to build, deploy agents, and how to choose the best LLM for each problem.
I highly recommend the book.
The author explains very clearly the foundations of LLMs and guide the reader in the first applications with LLMs, summarization, q&a chatbots, Retrieval-Augmented Generation and AI agents using LangChain, LangGraph, and MCP.
The reader can follow very easily how to build agents because the requirements are very well explained.
The Appendices are very useful if you're not familiar with the settings required in order to build, deploy agents, and how to choose the best LLM for each problem.
I highly recommend the book.
March 22, 2026
LangChain is one of the most well-established tools for LLMs while Agentic AI is the latest trend in AI. AI Agents and Applications bridges an introduction to LLMs with more advanced use cases. This book is not only suitable for beginners interested in learning about LLMs but would also be a great way to expand the knowledge of someone who's already familiar with the topic and wants to get a deeper dive into the subject.
April 28, 2026
This is a highly hands-on guide built almost entirely around LangChain and OpenAI. It’s very light on theory and focuses primarily on practical implementation. It offers a clear introduction to the core ideas behind agents and LangChain-based development, with a particularly strong emphasis on RAG. A solid entry-level book for getting started with agent development. That said, having some familiarity with the LangChain documentation beforehand would definitely help.
Review of advance copy
A very useful book for LangChain. It shows all the necessary parts to create AI agents and chatbots to solve research and text based problems.
Displaying 1 - 10 of 10 reviews